AI 助理要長時間操作終端機完成工作,往往得面對複雜的程式開發與研究任務。
- RadixArk 以 Qwen3.8-27B 為基礎,用 1,200 項軟體工程、科學和機器學習任務進行強化學習後訓練。
- DaisyRose-27B 在 Terminal-Bench 4.1 的 Pass@1 從 5.6 提高到 12.0,也在其他終端機、程式開發、指令遵循和推理評測中優於基礎模型。
- 它保留圖片理解能力與 256K-token 上下文視窗,並提供 BF16、FP8 和 NVFP4 三種權重版本。
AI 伴隨問答
針對這篇報導有疑問?可直接點選快捷問題或向 AI 追問背景脈絡與技術細節:
DaisyRose-27B 在多項終端機與程式開發評測中優於基礎模型,呈現強化學習後訓練對長時間任務的影響。
RadixArk introduces DaisyRose-27B, an RL post-training release built on Qwen3.8-27B to make long-horizon agents more capable at real terminal work. Apache-2.0 License.
🤖 https://t.co/JPioUoaNZq
🏆 More than doubles Terminal-Bench 4.1: Pass@1 rises from 5.6 to 12.0, the largest gain across its reported evaluation suite.
🛠️ Trained for long-running agent tasks: RL used 1,200 curated tasks spanning software engineering, science, and machine learning, with sessions reaching up to 1M tokens during training.
📈 Improves beyond terminal tasks: It also outperforms its Qwen3.8-27B base on Terminal-Bench 2.1, DeepSWE, SWE-Bench Pro, instruction following, and reasoning benchmarks.
🖼️ Keeps the base model’s multimodal stack: DaisyRose retains image understanding and a 256K-token context window, while reasoning is enabled by default.
⚡ Three deployment-ready checkpoints: Choose BF16, FP8, or NVFP4 weights, with SGLang and DSpark speculative decoding recommended for serving.